Abstract
The interpretation of (18)F-DOPA PET/CT performed for assessing nigrostriatal dysfunction (NSD) is usually based on visual assessment of the uptake in the basal ganglia (VA-BG). In the present study, we evaluate the diagnostic performance of an automated method that assesses BG uptake (AM-BG) and of methods that assess pineal body uptake, and examine whether these methods can enhance the diagnostic performance of VA-BG alone. We retrospectively included 112 scans performed in patients with clinically suspected NSD who also had a subsequent final clinical diagnosis provided by a movement disorder specialist (69 NSD and 43 non-NSD patients). All scans were categorized as positive or negative based on (1) VA-BG, (2) AM-BG, and (3) qualitative and semiquantitative assessment of pineal body uptake. VA-BG, AM-BG, assessment of pineal body (18)F-DOPA uptake by VA (uptake > background), by SUVmax (≥0.72), and by pineal to occipital ratio (POR ≥ 1.57) could all significantly differentiate NSD from non-NSD patients (Pv < 0.01 for all five methods). Of these methods, VA-BG provided the highest sensitivity (88.4%) and accuracy (90.2%). Combining VA-BG with AM-BG did not improve diagnostic accuracy. An interpretation algorithm that combines VA-BG with pineal body uptake assessment by POR calculation increased sensitivity to 98.5%, at the expense of decreased specificity. In conclusion, an automated method that assesses (18)F-DOPA uptake in the BG and assessment of pineal body (18)F-DOPA uptake can significantly separate NSD from non-NSD patients, with apparent inferior diagnostic performance when applied alone compared with VA-BG. When VA-BG categorizes a scan as negative or equivocal, assessment of the (18)F-DOPA uptake in the pineal body has the potential to minimize the rate of false negative reports. Further research is essential to validate this approach and to study the pathophysiologic relationship between (18)F-DOPA uptake in the pineal body and nigrostriatal dysfunction.